Improving WSD using ISR-WN with Relevant Semantic Trees and SemCor Senses Frequency

Proceedings of Recent Advances in Natural Language Processing, pages 233–239, Hissar, Bulgaria, 12-14 September 2011. Improving WSD using ISR-WN with Relevant Semantic Trees and SemCor Senses Frequency Yoan Gutiérrez Department of Informatics University of Matanzas, Cuba. {yoan.gutierrez}@umcc.cu Sonia Vázquez and Andrés Montoyo Department of Software and Computing Systems University of Alicante, Spain. {svazquez, montoyo}@dlsi.ua.es Abstract In this paper we concentrate on the resolution of the semantic ambiguity that arises when a given word has several meanings. This specific task is commonly referred to as Word Sense Disambiguation (WSD). We propose a method that obtains the appropriate senses from a multidimensional analysis (using Relevant Semantic Trees). Our method uses different resources WordNet, WordNet Domains, WordNet-Affects and SUMO, combined with senses frequency obtained from SemCor. Our hypothesis is that in WSD it is important to obtain the most frequent senses depending on the type of analyzed context to achieve better results. Finally, in order to evaluate and compare our results, it is presented a comprehensive study and experimental work using the Senseval-2 and Semeval-2 data set, demonstrating that our system obtains better results than other unsupervised systems. 1 Introduction The main goal of knowledge technologies is to provide meaning to the huge quantity of information that our multilingual societies generate day to day. A wide range of advanced techniques are required to progressively automate the knowledge lifecycle. For that, after performing an analysis to large data collections it is necessary to develop different approaches to automatically represent and manage a high-level of meaningful concepts (Montoyo et al., 2005). Moreover, to be able to create efficient systems of Natural Language Processing (NLP) it is necessary to turn the information extracted from words in plain text into a Concept Level or meaningful word senses. This representation allows to group words with similar meanings according to the context where they appear. In order to determine the right meanings of words in different contexts WSD systems have been developed. Furthermore, it has been proved that applications such as Machine Translation, Information Extraction, Question Answering, Information Retrieval, Text Classification, and Text Summarization require knowledge about word meanings to obtain better results. So, WSD is considered an essential task for all these applications (Ide and Véronis, 1998). For this reason many research groups are working on WSD, using a wide range of approaches. Due to the need of evaluating different approaches to show the improvements of NLP tasks it was created the Senseval1 competition. The first Senseval was in 1998 at Herstmonceux Castle, Succex (England) and after that every three years a new competition takes place. In Senseval, different NLP tasks are defined in order to evaluate systems using the same repositories and corpus. At present, the results obtained in WSD have been going poorer, because the requirements of each corpus are getting more complex. For example, in Senseval- 2 (Cotton et al., 2001) the best system obtained a 69% of accuracy in WSD, three years later in Senseval-3 (Snyder and Palmer, 2004) the best results were around …

Content truncated to 500 words